INDEX

Explanations

words related to misinformation, inaccuracies, and misconceptions

oai_token-act-pair · gpt-3.5-turbo

terms related to inaccuracies and misinformation

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

amen

-0.77

agra

-0.75

mun

-0.75

gans

-0.73

imen

-0.72

atom

-0.71

shall

-0.69

spring

-0.69

lining

-0.68

alm

-0.68

POSITIVE LOGITS

 misconceptions

1.08

 misinformation

1.07

 inaccur

1.06

 misconception

1.01

 inaccurate

1.00

 misunderstanding

0.96

 incorrectly

0.95

 misinterpret

0.92

 mistaken

0.92

 misunderstand

0.90

Activations Density 0.022%

words related to misinformation, inaccuracies, and misconceptions

terms related to inaccuracies and misinformation

No Comments

No Known Activations